AI Websites, Crawling and Search Console updates (Q2 ‘26)

[MUSIC PLAYING]
JOHN MUELLER: Hello, everyone.
And welcome to this episode of "Google Search News!"
I hope life is treating you reasonably well wherever
you are.
My name is John Mueller.
I'm your host today here from Google Switzerland.
Today, we have news about Search Console, AI-created websites,
crawling, Google Search and more.
Time is of essence, so let's dive right in.
Search Console launched a number of cool features
since we last met.
First up is a way to separate non-branded
from branded queries, meaning someone
was explicitly looking for your business.
In Search Console Insights, there's
a comparison between these query types.
The performance report shows more details.
The split is created with AI, based on multiple components.
It's not a regular expression.
These details are now available for all top-level sites
with enough traffic.
Then we have an AI-powered configuration tool
for the performance report.
This tool makes it easier to find the data
that you need for your day-to-day decisions,
especially if you're not as familiar with the report.
Also in the report, we added options
for weekly and monthly aggregations.
Sometimes the daily ups and downs
make it hard to spot trends or issues.
So hopefully, this makes your life a little bit easier.
Next, and currently only for a small set of sites,
there's an option to add search data for your company's
social profiles in the insights report.
We're still experimenting with this,
so it might be a while before it's more broadly available.
But if you see it, we'd love to get your feedback.
Please use the Thumbs Up or Thumbs Down buttons
next to the report to let us know what you think.
Finally, I mentioned query groups and custom annotations
last time.
Check them out if you haven't tried them yet.
I know the Search Console team has
some amazing next features lined up, so, as they say, stay tuned.
And now over to something that's not news from Google.
Let's talk briefly about so-called vibe-coded websites,
ones that were created primarily by AI tools.
These are essentially normal websites,
so they can be fine for Search.
However, I wanted to touch upon a few things that we noticed.
Most importantly, make sure that the content you're publishing
is actually adding value to the web.
It's easy to just create a site nowadays.
But if you want people to appreciate it,
make sure that it's actually good.
Then check out our SEO Starter Guide.
It's good background information on how Search works.
If your vibe-coded website has multiple pages,
double check the link rel="cannonical" elements.
Ideally, include the full URL of each page,
including your domain name.
If your website was built using a JavaScript framework,
like React or Next.js, understand
that things may be a bit more complicated
than with a static HTML website.
In particular, test your pages to make sure that Google
can view them properly.
Check out our JavaScript SEO documentation
and videos for more.
And one last thing-- add your site
to Search Console so that you're informed about potential issues
and the site's performance.
Again, for the most part, a vibe-coded website
will just work in Search.
The ones I've seen also have proper titles
and use structured data, which is all great.
If all this made you curious, I'd
recommend just trying it out.
You can create websites with normal AI systems,
like Gemini, but also with more dedicated tools,
like Antigravity or AI Studio.
Vibe-coded websites might not take over the popular web
just yet, but especially if you work with smaller clients,
it can be worth getting some practice with them.
And now over to the world of crawling.
We created a new site for general crawling-related
information, starting with some migrated Search content.
We added documentation about crawlers
for Read Aloud, NotebookLM, Pinpoint and Google Agent.
That last one, Google Agent, is used by AI agents
hosted on Google infrastructure, which you can use, too.
Also there, we created a high-level guide on things
to know about Google's web-crawling.
We occasionally get questions about crawling,
and this page tries to answer some of them.
It might be a useful reference to point others at.
Finally, we also updated the documentation
on the fetch limits for Googlebot.
The limit is 2 megabytes for uncompressed initial HTML,
but there are more details.
Most sites don't need to worry about this.
But if your site has giant menus in the HTML,
it's worth double-checking.
There's a blog post and a podcast episode for more.
And a few short updates from Search.
First, we had a Discover Corps update, a spam update,
as well as a normal Core update.
These updates are a part of how we work
to improve our Search results.
For e-commerce sites, Google Shopping
announced support for UCP, Universal Commerce Protocol.
UCP creates a common language for agents that interact
with your site and business.
This is still quite new.
It's not something all e-commerce sites
need to jump on just yet.
Now over to some posts from the SEO community.
I always find way too many articles that I want to pass on.
It's hard to pick just a few.
Here is a selection for you.
First, MJ Cachon's "Complete Guide to E-Commerce SEO."
It's just so comprehensive.
E-commerce can be quite complex.
I'll add links to all of these in the description.
Then, Dawn Anderson on demystifying
generative information retrieval,
clarifying some of the points of confusion
that you might have seen on social media.
Similarly, reflecting on the world around SEO and AI,
there's a great post by Lily Ray.
Both of these posts can be useful to push back
on some of the common misconceptions around SEO
and AI.
Finally, Amy Jurenka wrote a great guide
on the role of informational content in the age of LLMs.
There's no magic trick, but Amy highlights the value
of publishing content that adds value
to the web, my favorite kind of content.
Go check them out.
And, of course, keep making awesome things
that I can share with everyone.
And with that, just a few more short updates and we'll be done.
First up, Google Trends recently launched a new Explorer page.
It includes an AI feature to help
you discover terms to compare.
It's a great way of exploring what people are searching for
and, of course, just fun to try out.
Then there's Events.
We did Search Central live events in Brazil and Argentina
earlier this year.
It was great having folks meet with you in person.
Coming soon, we have events in Toronto, Canada;
Shanghai, China; and Sydney, Australia.
If you're in the region, check out our Events page
and register your interest for a ticket early on.
Finally, and perhaps particularly important
for computer users of a certain generation,
Bernd Wunsch created a wonderful robots.txt file, which is valid,
and includes a copy of the classic game "Doom."
Making robots.txt files is generally pretty easy.
You don't need to include games, but you can.
And with that, this episode of "Google Search News!"
is now complete.
Thank you for tuning in.
I hope this video was useful.
And please, add feedback and comments here.
We read them all.
If you subscribe to this channel,
we'll let you know when another episode is ready.
Bye.
[KEYBOARD KEYS CLATTERING]
[MUSIC PLAYING]